Note from GCF: I would disagree that the choice of C++ versus Java is
"obvious" for this is a PhD thesis not a deployment project. In my opinion
Java is likely to be preferred language for this class of applications in the future
and so it is important to discuss this point. PhD Theses are meant to
focus on research issues not optimal deployment today.
However I do think that choice you made : C++ with Java look and feel
was correct and is a strong point of thesis.

Referee Number 1

C++/Java:  I suggest that the choice of C++ be presented very simply.
There's no real question that the platform on which he worked, the need
for speed and access to full system resources justifies his choice of C++.

This may also eliminate the somewhat naive section on performance
measurement (for which a seriously optimizing compiler would optimize
away all the measurements).  If some of that is retained (and there is
some useful information) he should acknowledge that it's a "toy"
benchmark and pay attention to the fact that C++ optimized away his
function.

If some of the Java/C++ discussion remains, he should not say that
"Java doesn't have pointers".  It has lots of them, though not to
arbitrary locations.

Incidentally, I would say that pages 159 and 160 suggest that C++ isn't
so wonderful either, though I've seen much worse.


    From my viewpoint, a decent explanation of the "continuous execution"
facility is the most important missing ingredient.  There is a some
discussion of the problems involved, but not much explanation of the
solutions, unless one can dig them out of the appendices.  This probably
requires clarifying the relationships between the NeatTools module
framework, which doesn't seem to have much to do with concurrency, and
the concurrency introduced within the modules, of which there's a good
deal.  I suspect there are (informal) protocols for modules which
Yuh-Jye understands very well but hasn't articulated.  I see parts of
it in appendix 6, but extracting a clear understanding from the appendix
is not a task I wan't to undertake.

    Let me strongly second the request for editing.  It's not just
that the language is ungrammatical, inelegant, and awkward.  I'm often
not sure just what he intends to say.